Peter Robinson pbrobin...@gmail.com writes:
You can't really but then the repo is read only so there's no new
updates. We're currently building in koji and soon that should be in a
state where we can start doing daily composes like all other
architectures and have a minimal koji built image
in binutils, which has actually been
fixed already (the upcoming binutils release should contain this fix):
http://cygwin.com/ml/binutils/2013-07/msg00013.html
So the failure should disappear eventually with binutils rebase.
Please review.
Thanks,
PM
Petr Machata (2):
Support new 'h
Signed-off-by: Petr Machata pmach...@redhat.com
---
src/ChangeLog |5 +
src/readelf.c |3 +++
2 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/src/ChangeLog b/src/ChangeLog
index e538a57..49fc728 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,8 @@
+2013-09-30
the failure should disappear eventually with binutils rebase.
Please review.
Thanks,
PM
Petr Machata (2):
Support new 'h' core note item format for hidden fields
Support AArch64 architecture
backends/ChangeLog | 12 +
backends/Makefile.am| 17
Signed-off-by: Petr Machata pmach...@redhat.com
---
src/ChangeLog |5 +
src/readelf.c |3 +++
2 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/src/ChangeLog b/src/ChangeLog
index 51a5bd5..3beacd7 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,8 @@
+2013-11-14
Pushed.
Thanks,
PM
___
Ltrace-devel mailing list
Ltrace-devel@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/ltrace-devel
Adam Malinowski a.malinow...@partner.samsung.com writes:
Adding #include inttypes.h solves this problem.
I fixed this on master.
Thanks,
PM
___
Ltrace-devel mailing list
Ltrace-devel@lists.alioth.debian.org
Adam Malinowski a.malinow...@partner.samsung.com writes:
I've pulled fresh changes and see that arguments are displayed correctly.
Nice, thanks for testing it.
Thanks,
PM
___
Ltrace-devel mailing list
Ltrace-devel@lists.alioth.debian.org
Applied.
Thanks,
PM
___
Ltrace-devel mailing list
Ltrace-devel@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/ltrace-devel
Adam Malinowski a.malinow...@partner.samsung.com writes:
Seccomp is kernel tool which can filter system calls and perform
OK, thanks for that.
So what i did with ltrace? I added new setting to ptrace which enables
seccomp events, also I added seccomp events handling and added
I imagine
I put this on master. Thanks for going through with this.
PM
___
Ltrace-devel mailing list
Ltrace-devel@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/ltrace-devel
Hi there,
I have a bunch of patches to implement tracing of IRELATIVE PLT slots
and STT_GNU_IFUNC symbols. It all sits on pmachata/ifunc.
IFUNC symbols are used for embedding code snippets that decide what
actual symbol should a call resolve to. If a PLT slot resolves to an
IFUNC symbol,
Mark Wielaard m...@redhat.com writes:
diff --git a/libdw/cfi.c b/libdw/cfi.c
index b1e8d48..67e3dfb 100644
--- a/libdw/cfi.c
+++ b/libdw/cfi.c
@@ -252,6 +252,8 @@ execute_cfi (Dwarf_CFI *cache,
continue;
case DW_CFA_expression:
+ /* Expression rule relies on section
Roland McGrath rol...@hack.frob.com writes:
I'm not very happy about readelf encoding all this format knowledge.
That's what we have libebl for. If the Ebl_Core_Item format needs to be
extended to handle these new note types, then that is what we should be
doing rather than hard-coding
FYI, I updated elf.h from libc again, this time mainly for the AArch64
bits.
PM
8--
Signed-off-by: Petr Machata pmach...@redhat.com
---
libelf/ChangeLog |4 ++
libelf/elf.h | 152
Luca Clementi luca.cleme...@gmail.com writes:
For each stack frame this patch prints the library path containing the
code pointed by the IP.
The output format is similar to the return value of backtrace_symbols
function found in glibc, e.g.:
./a.out() [0x40063d]
./a.out() [0x4006bb]
Signed-off-by: Petr Machata pmach...@redhat.com
---
The use case here is NT_ARM_HW_BREAK and NT_ARM_HW_WATCH core notes.
These end in a 4-byte padding field. If that field is absent from the
note description, readelf notices that we still have data left, and
does another pass through
Mark Wielaard m...@redhat.com writes:
+return printf (buffer overrun), 0;
ehe, OK, I admit I didn't know you could do that with a return statement.
I assume it returns zero?
Yeah, that's just return EXPRESSION where the expression involves a
comma operator. Nothing magic about it. I
Signed-off-by: Petr Machata pmach...@redhat.com
---
src/ChangeLog |6 ++
src/readelf.c | 103 +++
tests/ChangeLog |6 ++
tests/Makefile.am |2 +-
tests/run-readelf-mixed
Signed-off-by: Petr Machata pmach...@redhat.com
---
src/ChangeLog |5 +++
src/readelf.c | 67 --
tests/ChangeLog |5 +++
tests/run-readelf-mixed-corenote.sh | 11 ++
4 files changed, 84
That's all fine, I put it all on master. Patches for the typedef
warning and free in printf are suiteble for 0.7 as well, so I put them
there. A test case for the latter would be nice, could you contribute
one, please?
Thank you,
PM
___
Ltrace-devel
Petr Machata pmach...@redhat.com writes:
I haven't had a chance to look at NT_FILE yet, but I intend to.
pmachata/NT_SIGINFO now includes code for handling NT_FILE as well.
I changed the READ_INT etc. macros to functions (they can be reused for
NT_FILE) and rewrote the branch history
Mark Wielaard m...@redhat.com writes:
I guess it makes sense to release early and release often. I was hoping
to get some more stuff in (unwinder and arm64 backend) before the next
release. But those will probably take some time to polish.
Oh, but we can make _another_ release when those come
Hi there,
I noticed that newer kernels put NT_SIGINFO and NT_FILE core notes into
core dumps.
NT_SIGINFO is tricky as interpretation of fields further in the note
depends of those earlier in. I don't think our core note item mechanism
can handle this. Instead, I made them special, they are
Mike Frysinger vap...@gentoo.org writes:
Trying to build with clang fails due to the use of a small nested
function. Change it to a macro instead.
Please change it to a function instead.
Thanks,
PM
When a library is mapped in, ltrace looks for soname.conf
oh, I was not aware of this feature (and do not find any documentation
about it).
It's now documented it in ltrace.1.
Thanks,
PM
___
Ltrace-devel mailing list
Enrico Scholz enrico.sch...@sigma-chemnitz.de writes:
Due to '-I.', the local 'glob.h' makes it impossible to include the
system wide glob.h header. Patch renames 'glob.h' to avoid this
conflict.
Ugh, you are right, that's not very nice. But would you please rename
it to ltrace-glob.h
Enrico Scholz enrico.sch...@sigma-chemnitz.de writes:
Patch adds an 'include' statement which can be used to include other
I do agree an include mechanism is necessary, but I envision something a
bit different than you.
configuration files recursively. This makes it e.g. possible for
linux
Ralf Corsepius rc040...@freenet.de writes:
However, I recall another case, deps on dirs often don't work:
(massively) parallel make.
Did you try to build the package single-threaded (make -j1)?
No, it used -j2 (as bug 885474 suggests).
PM.
--
devel mailing list
URL:
http://savannah.gnu.org/bugs/?39934
Summary: make closes a wrong FD, results in *** INTERNAL:
readdir: Bad file descriptor
Project: make
Submitted by: pmachata
Submitted on: Втр 03 Сен 2013 17:44:02
Severity: 3 -
Follow-up Comment #2, bug #39934 (project make):
This is consistently reproducible with current git HEAD (40a49f24) as well as
3.82.
___
Reply to this item at:
http://savannah.gnu.org/bugs/?39934
Ralf Corsepius rc040...@freenet.de writes:
I guess, no. AFAIS, this makefile carries deps on directories.
This is a very old known general limitation of and portability isse
with make and one of known donts.
Deps on dirs work on local Linux file systems, but doesn't work on
linux nfs and
Markos Chandras hwoar...@gentoo.org writes:
/bin/sh: line 17: cd: mips: No such file or directory
This is because the mips-mipsel symlink (in sysdeps/linux-gnu) does
not exist in the tarball
We forgot to add that to EXTRA_DIST in sysdeps/linux-gnu/Makefile.am.
I added fix to 0.7 branch. Can
URL:
http://savannah.gnu.org/bugs/?39851
Summary: Stack limit not restored for processes spawned
through $(shell)
Project: make
Submitted by: pmachata
Submitted on: Чтв 22 Авг 2013 14:54:01
Severity: 3 - Normal
Follow-up Comment #1, bug #39851 (project make):
Hmm, this breaks targets/SECONDARY.
___
Reply to this item at:
http://savannah.gnu.org/bugs/?39851
___
Message sent via/by Savannah
Follow-up Comment #2, bug #39851 (project make):
Scratch that, that seems to be interaction between this and other downstream
patches. The upstream test suite consistently passes.
___
Reply to this item at:
Follow-up Comment #3, bug #39851 (project make):
OK, the test is racy. I see it was fixed in 3057357c from 10 Dec 2011.
___
Reply to this item at:
http://savannah.gnu.org/bugs/?39851
___
Peter Robinson pbrobin...@gmail.com writes:
Maybe we need to put a mass rebuild starts point in the Schedule in
the future so that people are more aware of this and have the sorts of
features like a perl rebase done in reasonable time.
That would be useful.
Thanks,
PM
--
devel mailing list
Ahmad Samir ahmadsamir3...@gmail.com writes:
$ yum list boost-static*
[...]
Available Packages
boost-static.i686
1.53.0-6.fc19fedora
boost-static.x86_64
1.53.0-8.fc19updates
it looks like
punto...@libero.it punto...@libero.it writes:
there is also zookeeper... to be rebuild with the new boost?
Yeah, I only got around to rebuilding those that directly depend on
Boost DSO's. I guess I can order builds of the rest of the dependencies
today, though originally my plan was to go
More failures from another bunch of rebuilds of about 100 packages that
have API-only dependence on Boost. I fixed those overtly Boost-related,
what remains seems to fail due to something else, so it should be OK to
just fix these in Rawhide and ignore Boost.
Note that f20-boost will be merged
Hi there,
as some of you may have noticed, the biannual Boost rebuild has been
underway since Saturday! So far about 100 packages have been rebuilt.
I'll appreciate any help that I can get with resolving the current
failures. Just ping me on IRC (_petr) so that we don't duplicate
effort.
Petr Machata pmach...@redhat.com writes:
as some of you may have noticed, the biannual Boost rebuild has been
underway since Saturday! [...] I'll appreciate any help that I can
get with resolving the current failures.
I forgot to mention that if you wish to build Boost clients, you should
Ville Skyttä ville.sky...@iki.fi writes:
These are now fixed in master.
Thanks!
PM
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
punto...@libero.it punto...@libero.it writes:
Il 29/07/2013 18:01, Petr Machata ha scritto:
bookkeeper 5663346 Package: xbean-3.13-2.fc20.noarch (build)
Requires: eclipse-equinox-osgi
sorry, i rebuilt without boost 1.54.x support...
eclipse-equinox-osgi is available also in arm
Nathan Ridge zeratul...@hotmail.com writes:
On 3 July 2013 02:41, Nathan Ridge wrote:
./boost/bind/arg.hpp:37:22: warning: typedef ‘T_must_be_placeholder’
locally
defined but not used [-Wunused-local-typedefs]
It was new in GCC 4.7, http://gcc.gnu.org/gcc-4.7/changes.html
I've seen spews
Dan Mashal dan.mas...@gmail.com writes:
The Feature page still belongs to the FeaturePageIncomplete category.
I feel a somewhat similar way about boost and it would be nice if
there was some more detailed descriptions here.
Frankly, my biannual filing of Feature/Change page is mostly
Petr Machata pmach...@redhat.com writes:
Dan Mashal dan.mas...@gmail.com writes:
The Feature page still belongs to the FeaturePageIncomplete category.
I feel a somewhat similar way about boost and it would be nice if
there was some more detailed descriptions here.
So if the consensus
Petr Pisar ppi...@redhat.com writes:
Is the installed libSDL.so symlink a mistage in the SDL-devel package?
Is renaming libSDL.so to libSDL-1.2.so wise? The libSDL.so is used in
upstream and other distributions.
I'm speculating here, but renaming the actual DSO like this would make
it
Richard Shaw hobbes1...@gmail.com writes:
I initially got abi-compliance-checker into Fedora because one of my
packages does not maintain any sort of API/ABI compatibility or even
versioning for that matter. That way I could always check a new
release to see if any of its dependencies needed
Summary of changes:
a4b73be... Rebuild for TBB memory barrier bug (*)
(*) This commit already existed in another branch; no separate mail sent
--
Fedora Extras Perl SIG
http://www.fedoraproject.org/wiki/Extras/SIGs/Perl
perl-devel mailing list
perl-devel@lists.fedoraproject.org
This was long overdue, last update was almost a year ago. That said,
the update should be safe: upstream-tracker.org lists only one warning
that is potentially ABI-breaking, which is that the constant eid_max
changed value. That constant doesn't appear to be used by any of the
clients. Soname
Akim Demaille a...@lrde.epita.fr writes:
Le 17 mai 2013 à 17:28, Petr Machata pmach...@redhat.com a écrit :
several discrepancies have been found between what command-line options
Bison handles and what's documented[1]. Looking in, I found that most
of the options are actually a dead code
in f6bd5427 on
2001-11-26 as a %-directive, with comments that it's not yet
implemented. It was converted to a command-line option later, but
doesn't seem to ever have been actually implemented.
Signed-off-by: Petr Machata pmach...@redhat.com
---
src/getargs.c |9 -
1 files changed, 0
Mark Wielaard m...@redhat.com writes:
Make sure the highest address for the CU is marked as end_sequence.
This is required by the DWARF spec, but some compilers forget and
dwfl_module_getsrc depends on it. We could reject it as bad DWARF
but the DWARF .debug_lines spec is much stricter than
Markos Chandras markos.chand...@gmail.com writes:
On 21 March 2013 16:25, Petr Machata pmach...@redhat.com wrote:
Finally, would you please also update README with triplets of systems
where you think ltrace should work with this patch?
The patch already adds this in the README file:
metag
Markos Chandras markos.chand...@gmail.com writes:
This patchset adds support for Imagination's Meta architecture.
The Meta Linux kernel port will be included in the Linux Kernel
v3.9. It also uses the generic system call numbers.
Pushed.
Thanks,
PM
Mark Wielaard m...@redhat.com writes:
I am undecided whether to fix the dwarf_getaranges code to do the same
(in normal cases the length and terminator for the table match up, so it
shouldn't matter).
I still like to add this raw .debug_aranges reader in readelf because
I think it is useful
Joe Damato ice...@gmail.com writes:
On Tue, Mar 19, 2013 at 6:25 PM, Petr Machata pmach...@redhat.com wrote:
Joe Damato ice...@gmail.com writes:
joe@ubuntu:~/code/ltrace$ /tmp/test/usr/local/bin/ltrace -e* -p 2821
libm.so.6-(0x7fff07366f90, 0x7fff07366f90, 0, -1
Joe Damato ice...@gmail.com writes:
I noticed that ltrace was unable to trace symbols from runtime loaded
dynamic objects when attaching to a PID. I've written a small (and
ugly) patch to try to fix this. Let met know if there is a better way
to do this and I'll be happy to make the change
Dan Horák d...@danny.cz writes:
Josh Stone píše v Út 05. 03. 2013 v 09:44 -0800:
Is that feasible for C++ APIs? I mean, it might be possible if you're
*really* careful about hiding class changes, but this project is not
structured that way.
it is, see eg. the wxWidgets library, they are
Oliver Spornitz ospor...@rz-online.de writes:
Am 16.02.2013 um 15:58 schrieb Oliver Spornitz:
I applied the three patches to the branch ospornit/move_mipsel on top
of the current master.
I missed to move mipsel.h to mips.h in the third commit. I added this
move to that commit an recreated
Sedat Dilek sedat.di...@gmail.com writes:
with latest ltrace upstream GIT I get for my 7170 (mipsel-based) router box:
configure: creating ./config.status
config.status: error: cannot find input file: `Makefile.in'
As a wild guess: did you run ./autogen.sh, make distclean, etc.?
I.e. isn't
Andrey Zonov z...@freebsd.org writes:
I don't see a reason to keep this work in separate branch because it
doesn't touch any other code. I have 3 patches that add FreeBSD/x86
support, I can send them if you like to review.
If you want to have them in official ltrace, then please send them.
There seems to be something wrong with either your diff program, or your
mailer, or alternatively something mangled it on the way. Maybe try to
post as attachment? I generally prefer inline, but if it's broken...
Thanks,
PM
Andrey Zonov z...@freebsd.org writes:
diff --git a/lens_default.c
Kevin Fenzi ke...@scrye.com writes:
On Fri, 08 Feb 2013 02:57:19 +0100
Petr Machata pmach...@redhat.com wrote:
Hi there,
I have just built boost 1.53. I didn't go through the side tag as
originally envisioned, as tomorrow's mass rebuild should take care of
it all in one fell swoop
Andrey Zonov z...@freebsd.org writes:
When do you plan to release 0.8.0? I just ask because FreeBSD/x86 port
is ready. It would be great to get it in 0.8.0.
There's no fixed date, but soon. End of February? There's no real
hurry, we can wait with 0.8.0 until you get the patches in. I'll
Sedat Dilek sedat.di...@gmail.com writes:
is it possible to get a v0.7.3 - not only for MIPS folks?
Hmm, I think it might make sense... we now support big endian MIPS, and
there's a couple more nits. I'll do it together with 0.8.0.
Regarding that, what should really be in README? Up until
Andrey Zonov z...@freebsd.org writes:
diff --git a/glob.c b/glob.c
index b26637f..06fec47 100644
--- a/glob.c
+++ b/glob.c
@@ -27,7 +27,7 @@
static ssize_t
match_character_class(const char *glob, size_t length, size_t from)
{
- size_t i;
+ size_t i = 0;
if (length 0)
Jerry James loganje...@gmail.com writes:
Upstream has one problem, though. On the other supported platforms
(Debian, various flavors of BSD), the reloc type defines
(R_ARM_MOVT_ABS, etc.) are available in some userspace header. I see
them in the Llinux kernel sources, in
Kevin Fenzi ke...@scrye.com writes:
I know the last cycle the boost tag was merged back in and there were a
lot of packages that still needed rebuilding. For this cycle, do you
have any provenpackagers in your feature owners that could just make
sure all the packages are rebuilt? It would be
Hi there,
as every release, we (the Boost maintainers) intend to rebase Boost for
Fedora 19. The targeted release is 1.53.0. The plan is outlined on the
feature page [1].
Boost 1.53 is not out yet (it will be on February 4). Beta is out, but
I don't think it's a good idea to rebase to that.
Hello,
the tarballs released for 0.7.0 were broken--they weren't made from a
clean checkout, and contained some work-in-progress files from my tree.
I've now corrected the mistake. As previously, new tarballs were
uploaded to:
https://alioth.debian.org/frs/?group_id=30892
The checksums of
Sedat Dilek sedat.di...@gmail.com writes:
One more thoughts on broken or golden toolchain... gcc's optimization-level.
The Freetz router project uses aggressive -Os by default... I could
try more commonly used -O2.
You are right, it is the -Os that causes this, and it hits on final gcc
as
Andreas Schwab sch...@linux-m68k.org writes:
---
sysdeps/linux-gnu/ppc/Makefile.am | 1 +
1 file changed, 1 insertion(+)
Applied. As I expected, 0.7.1 is due soon for these sorts of nits that
I missed.
Thank you,
PM
___
Ltrace-devel mailing list
Petr Machata pmach...@redhat.com writes:
diff --git a/sysdeps/linux-gnu/mipsel/plt.c b/sysdeps/linux-gnu/mipsel/plt.c
index 7799dfa..7bb00e0 100644
--- a/sysdeps/linux-gnu/mipsel/plt.c
+++ b/sysdeps/linux-gnu/mipsel/plt.c
@@ -251,6 +251,13 @@ void arch_symbol_ret(struct Process *proc, struct
Sedat Dilek sedat.di...@gmail.com writes:
ltrace-0.7.0/sysdeps/linux-gnu/trace.c:908:#if defined __sparc__ ||
defined __ia64___
3x _ for IA64 arch?
Right, there's no such thing as __ia64___. The code works anyway,
apparently ia64 either has hardware singlestepping, or it's emulated in
the
Andreas Schwab sch...@linux-m68k.org writes:
I'm getting a lot of testsuite failures due to
Couldn't determine base address of linux-vdso32.so.1
Couldn't load ELF object linux-vdso32.so.1: No such file or directory
What architecture is this on? Apparently l_name of VDSO in r_debug
linkmap
Andreas Schwab sch...@linux-m68k.org writes:
$ ltrace true
/etc/ltrace.conf:442: error: unknown type around '{
'
/etc/ltrace.conf:443: error: syntax error
/etc/ltrace.conf:444: error: unknown type around 'suseconds_t tv_usec;/*
microseconds */
'
/etc/ltrace.conf:445: error: unknown
Sedat Dilek sedat.di...@gmail.com writes:
On Sat, Nov 10, 2012 at 12:15 AM, Petr Machata pmach...@redhat.com wrote:
Petr Machata pmach...@redhat.com writes:
diff --git a/sysdeps/linux-gnu/mipsel/plt.c b/sysdeps/linux-gnu/mipsel/plt.c
index 7799dfa..7bb00e0 100644
--- a/sysdeps/linux-gnu
Hi there,
the test suite currently cleanly passes on x86's, ppc's, ia64 and
s390's, including 64/3{2,1} cross-tracing scenarios. On x86's, where
valgrind supports ptrace system calls, the test suite was verified to be
leak- and memory error free. On ppc's I made sure that tracing binaries
with
Edgar E. Iglesias edgar.igles...@gmail.com writes:
On Fri, Nov 02, 2012 at 06:07:50PM +0100, Sedat Dilek wrote:
This happens with latest ltrace-git...
./ltrace: proc.c: 755: breakpoint_for_symbol: Assertion `bp-libsym ==
((void *)0)' failed.
I've seen this bug, and I'm actually inclined
We got this in Fedora as well. There, it's easily reproducible by issuing
chrpath --hep. Full details here:
https://bugzilla.redhat.com/show_bug.cgi?id=868611
** Bug watch added: Red Hat Bugzilla #868611
https://bugzilla.redhat.com/show_bug.cgi?id=868611
** Patch added:
edgar.igles...@gmail.com writes:
From: Edgar E. Iglesias ed...@axis.com
When mapping function parameters into argument value dictionaries,
make sure to keep a 1-to-1 mapping between indexes.
For STOP parameters, we insert a ARGTYPE_NONE value to fill out
the argument value dictionary.
edgar.igles...@gmail.com writes:
From: Edgar E. Iglesias ed...@axis.com
Remove the check for generic tags allowing the caller
to pass extension tags. I couldn't find a way to strictly
validate the tag nr, furthermore the seek tag is internally
generated so I simply dropped the check (beyond
edgar.igles...@gmail.com writes:
+/* Are we in pure CPIC mode (the non-PIC ABI extension)? */
+static inline int mips_elf_is_cpic(unsigned int elf_flags)
Function name shoud start a new line.
+ /* FIXME: for CPIC we should really scan both GOT tables
+ * to pick up relocations to
Edgar E. Iglesias edgar.igles...@gmail.com writes:
On Thu, Oct 11, 2012 at 04:44:15PM +0800, Jackie Yeh [葉龍泉] wrote:
Your patches works!
I have to apply patch 1~6, only 1~4 doesn't work.
One question for the patches:
For these patches I copy-and-paste, then save to a patch file, but most
Jan Kratochvil jan.kratoch...@redhat.com writes:
On Tue, 09 Oct 2012 22:50:20 +0200, Petr Machata wrote:
Mark Wielaard m...@redhat.com writes:
Do you happen to know where
DW_OP_GNU_reinterpret is described, I assumed it was part of
http://www.dwarfstd.org/doc/040408.1.html but it isn't
Jackie Yeh [葉龍泉] jackieontra...@gmail.com writes:
Hi Petr,
I applied your patch, and attached the log with -D77 option.
:: fetch_rd32: couldn't read 20 bytes from (nil)
It's strange that 'addr' is a NULL pointer!
Indeed. It loads 0 from DT_DEBUG. What libc are you using? It seems
like
-10-09 Petr Machata pmach...@redhat.com
+
+ * dwarf_getlocation.c (__libdw_intern_expression): Handle
+ DW_OP_GNU_parameter_ref, DW_OP_GNU_convert, DW_OP_GNU_reinterpret,
+ DW_OP_GNU_regval_type, DW_OP_GNU_entry_value,
+ DW_OP_GNU_deref_type, DW_OP_GNU_const_type.
+
2012-08
Hi there,
on branch pmachata/valgrind are two testsuite fixlets and a patch for
running testsuite under valgrind. This is enabled by --enable-valgrind
passed to configure. The code was essentially cut'n'pasted from
elfutils, original author is Mark Wielaard. I consider this
uncontroversial and
edgar.igles...@gmail.com writes:
vect_reserve(struct vect *vec, size_t count)
{
if (count vec-allocated) {
- size_t na = vec-allocated != 0 ? 2 * vec-allocated : 4;
+ /* Allocate 4 extra slots for growth. */
+ size_t na = count + 4;
edgar.igles...@gmail.com writes:
From: Edgar E. Iglesias ed...@axis.com
Makes it easier to add new types.
Hmm, this was deliberate, but now that there are more back ends alive
than I can hope to take care of personally, your change makes sense.
Any new additions (like 128-bit doubles, vector
Edgar E. Iglesias edgar.igles...@gmail.com writes:
The issue is that if count is large enough, na might not end up as
count triggering the assert. We can keep the exponential growth with
another fix. I didnt realize that was important here.
Ah, I see. I'm hitting this problem with your
Sedat Dilek sedat.di...@gmail.com writes:
Ltracing the dsld-daemon (closed source) on my router box lead
sometimes to stalls and/or my box rebooted. This could be due to
limited RAM as I have approx. 2MiB free.
It is very possible that we leak stuff in ltrace. I'm currently working
on
edgar.igles...@gmail.com writes:
--- a/sysdeps/linux-gnu/mipsel/trace.c
+++ b/sysdeps/linux-gnu/mipsel/trace.c
@@ -202,7 +202,14 @@ int mips_next_pcs(struct Process *proc, uint32_t pc,
uint32_t *newpc)
newpc[nr++] = rx + ((pc + 4) ~0x0fff);
Edgar E. Iglesias edgar.igles...@gmail.com writes:
The patches look good. I've tried to replace the TOPLT_GOTONLY approach
used by MIPS with delayed syms and it works nicely.
Great, thanks. I'll merge after I get your MIPS optimizations in.
The patch looks fine content-wise, but what you sent
Edgar E. Iglesias edgar.igles...@gmail.com writes:
On Thu, Sep 27, 2012 at 01:36:48AM +0200, Petr Machata wrote:
edgar.igles...@gmail.com writes:
+#define ARCH_HAVE_ATOMIC_SINGLESTEP
I think I should rename this to ARCH_HAVE_SOFT_SINGLESTEP or some such.
The original intention
edgar.igles...@gmail.com writes:
+int
+arch_atomic_singlestep(struct Process *proc, struct breakpoint *sbp,
+int (*add_cb)(void *addr, void *data),
+void *add_cb_data)
+{
+ uint32_t pc = (uint32_t) get_instruction_pointer(proc);
+ uint32_t
edgar.igles...@gmail.com writes:
+/* When functions return we check if the symbol needs an updated
+ breakpoint with the resolved address. */
+void arch_symbol_ret(struct Process *proc, struct library_symbol *libsym)
+{
[...]
+
+ bp = malloc(sizeof (*bp));
+ if (bp == NULL)
801 - 900 of 1131 matches
Mail list logo